For years, brands have relied on an arsenal of tools to hear customer feedback. Surveys. Social listening. Focus groups. Ethnographic studies. From those and many other sources, businesses have been able to build a comprehensive understanding of customer wants and needs, enabling them to create more effective products and marketing strategies.

Now, businesses are entering a new frontier with AI. The adoption of AI (including generative AI) opens up new ways for companies to understand customers by applying advanced data analysis, real-time insights, and personalized interactions. It’s all very exciting – but not without risk.

Let’s take a closer look.   

The Always-on Persona   

At the heart of a typical marketing campaign is the customer persona: a detailed, semi-fictional representation of a business’s ideal customer based on market research and real-world data about existing customers. AI turns these personas into always-on personas, which are dynamic representations that continuously evolve based on real-time data and interactions.   

Always-on personas are special because they provide continuous, real-time insights and feedback, allowing brands to adapt rapidly to changing consumer preferences and market conditions. These AI-generated personas are constantly updated with new data inputs, ensuring they reflect the most current consumer behaviors and preferences.    

P&G employs AI-driven personas to understand consumer behavior better and personalize marketing campaigns. This helps the company tailor its advertising and product offerings to meet the specific needs and preferences of different customer segments.   

Real-Time Learning   

LEGO famously collaborates with customers by giving them a sandbox to share ideas for new designs. Generative AI makes that process faster and more iterative. Coca-Cola’s Real Magic initiative uses consumer-facing generative AI to enhance its marketing efforts. The company partnered with OpenAI to create a custom site where users can generate AI-infused images using Coca-Cola’s iconic brand elements. This technology, built on OpenAI’s DALL-E 3 image creation platform, was trained to understand and replicate Coca-Cola’s distinct features, such as its signature script and specific shade of red. Through this interactive platform, consumers input prompts to receive personalized images, effectively turning every user interaction into a learning opportunity for the AI model. This continuous engagement allows Coca-Cola to use the AI as a vast, ongoing focus group, gathering valuable insights and refining its marketing strategies based on real-time consumer feedback.  

Coca-Cola is not alone. Many other businesses are leaning into AI to learn from customer feedback in real time. For example, Stitch Fix uses generative AI to assist its stylists in interpreting customer feedback and providing personalized product recommendations. This application of AI helps the company better understand and meet individual customer preferences, thereby improving the overall customer experience.   

Hyper-Personalization   

AI enables businesses to process and act upon vast amounts of customer data in real-time, creating highly personalized experiences for millions of customers simultaneously. This level of hyper-personalization at scale, with continuous optimization and cross-channel consistency, is simply not feasible without the power of AI.  

For instance, Michaels Stores has used generative AI to enhance customer engagement through hyper-personalization. The company developed a content generation and decision-making platform that significantly increased the personalization of their email campaigns from 20% to 95%. This change led to a 41% increase in click-through rates for SMS campaigns and a 25% increase for email campaigns.  

Learning From Customer Service  

Chatbots are more than customer service tools. With AI, they provide a 24/7 customer feedback loop that businesses can use to become more responsive. With GenAI, chatbots can understand their customers by using language understanding, sentiment analysis, personalization, and predictive analytics. Every interaction with a generative AI chatbot provides valuable data on customer preferences, pain points, and behavior patterns. By analyzing this data, businesses can gain deeper insights into their target audience, refine their marketing strategies, and develop more effective products and services.  

For instance, Delta Airlines’ “Ask Delta” virtual assistant exemplifies how generative AI chatbots gather valuable data from every interaction. By analyzing customer questions, preferences, and pain points, Delta gains insights into its target audience, enabling the company to refine marketing strategies, improve customer experiences, and develop products and services that better meet customer needs. This continuous feedback loop demonstrates the power of generative AI chatbots in driving data-driven decision-making and improving customer satisfaction in the airline industry.   

A Caveat 

Here’s the catch: to realize the benefits of AI, you’ll probably need to build your own in-house large language model. The off-the-shelf tools such as ChatGPT have a fundamental drawback: they are running out of publicly available data to train themselves to get better. Epoch AI estimates that the stock of human-generated public text, around 300 trillion tokens, might be fully utilized between 2026 and 2032. This timeline could be accelerated if models are intensely overtrained.    

Companies are exploring the use of private data sources, such as proprietary documents and internal communications, to train specialized LLMs. This approach is less reliant on publicly available data and can provide unique insights specific to certain industries or organizations. And to do that, they’re spending millions of dollars to develop each model.   

But the upside to developing your own marketing LLM? You tap into your private first-party data to create more valuable insights than you can get from an off-the-shelf tool. Invest wisely.